KAM: an economic mechanism against white-collar AI displacement

Sydney, Australia

Raisingactive3 people

Selma: The AI that never leaves the building.

Project Details

At EthicAI we have built Selma, an on-premises AI that gives organisations a sovereign, ethically governed alternative to cloud dependency. We are now focusing on development of KAM: turning the attribution mechanic from an internal working model into a documented, IP-protected, system that tracks ongoing value. Our team consists of Co-founders, Kara Bombell: Operations, Katriel Healy: Org Design, and Simon Brock: Product. During this project we will formalise the methodology, file for IP protection on the mechanism, and test the full reimbursement pathway against at least one real staff departure. Upon completion, we will publish a public paper on the mechanism and what we learned.

Theory of Impact

Most AI x-risk framing centres on loss of control over powerful systems. We think there's a slower risk running underneath that: the wholesale absorption of professional expertise into AI systems with no mechanism to attribute or compensate the humans who generated it. Rapid, uncompensated displacement of white-collar work will cause unemployment. It erodes the tax base and the professional class that functioning institutions depend on for stability, at precisely the moment those institutions need to be capable of governing transformative AI well. A destabilised society with increasing wealth inequality governs AI badly.

We are working to build answer at organisational scale. The Knowledge Attribution Model tracks what each person contributes to a growing AI capability, weighted by impact, and gives them ongoing attribution for what the model generates from that contribution afterward. It rewards novel research and depth of expertise rather than volume of knowledge handed over: generating new understanding earns more than surrendering what's already known. Where the default path is to extract expertise into a model and discard the person, KAM keeps a durable, ongoing economic link between the human source of knowledge and the value it keeps generating.

People

Kara Bombell

Team Member

Funding Details

Private

Only visible to verified funders, reviewers, and admins.

Email hi@grantmaking.ai to get verified

Grants Received– no grants recorded

Add grant

Funding Asks

grantmaking.ai Launch Round

Applied

Minimum

$****

Ideal

$****

How the money will be spent

At EthicAI we have built Selma. This grant funds our KAM development: turning the attribution mechanic from an internal working model into a documented, IP-protected, system that tracks ongoing value. The minimum and ideal amounts would support this project at all levels providing funds to drive our next phase to completion.

Over six months we will formalise the methodology that determines how contribution to Selma's knowledge base is measured and valued, calibrated to reward novel research and depth of expertise over volume of data handed over. We will build the tracking layer that links downstream model outputs back to the contributions that produced them, so attribution continues after the knowledge is absorbed rather than stopping at the point of contribution. We will file IP protection on the mechanism, and test the full reimbursement pathway against at least one real staff departure, so someone who has left the organisation continues to draw value from a model still generating output from what they contributed. We will close the six months by publishing a public paper on the mechanism and what we learned, so other organisations have something to adopt rather than defaulting to extraction.

Success looks like an ongoing reimbursement running for at least one person who has left the organisation, and a published methodology other organisations can point to when deciding how to handle attribution as AI absorbs their experts' knowledge.

Discussion

No comments yet. Be the first to share your thoughts.